The present application is related to U.S. patent application Ser. No. 09/266,961 filed on Mar. 12, 1999, which is hereby incorporated herein by reference.
The present invention relates, in general, to methods for processing signals from nuclear uptake probes for radiation detection and, more particularly, to methods of adjusting a control unit to calibrate a radiation detection system each time a new probe is used.
Radioactive pharmaceuticals used in combination with radiation detection systems have been proven to be effective in locating radio labeled tissue within patients. These pharmaceuticals are also known as radionucleides and include solutions of Iodine 125, Iodine 131, and Phosphorous 32. Other radionucleides include monoclonal antibodies, peptides, and certain colloids labeled with radioactive isotopes such as Technetium-99. Once a radionucleide is introduced into a patient""s body, it will tend to collect at targeted tissue sites, such as, for example, lymph node sites and such sites may be located by looking for concentrations of the radionucleide.
The mammalian lymphatic system has various interrelated functions, including circulating and modifying tissue fluid formed in capillary beds and removing cell debris and foreign matter. For certain cancers, neoplastic cells migrate and collect at regional nodes within an associated lymph drainage basin. Some cancers, such as those encountered in the breast, will evidence somewhat predictable nodal involvement. The axillary lymph node region is the principal site of regional metastasis from carcinoma of the breast, and approximately 40% of patients have evidence of spread to the axillary nodes. In some approaches to the disease, these axillary nodes are removed as a form of therapy.
Sentinel node biopsy is a less invasive alternative to lymph node dissection in diagnosing metastasis of breast cancer tumors. The principle of sentinel node biopsy is that neoplastic cells detaching from the primary tumor are most likely to be held by the sentinel node, which is the first lymph node to receive lymph from the involved area and the most likely site of early metastasis. If the sentinel node is free of cancer, it is highly probable that all of the other nodes are free of cancer cells. This knowledge helps the physician in staging the disease.
Thus, it is important to identify the sentinel node when trying to determine whether cancer has metastasized. Detection of a sentinel node may be achieved by using a gamma ray detection probe intra-operatively to assist surgeons in locating tissue tagged with a radionucleide. U.S. Pat. No. 5,732,704 to Thurston et al discloses a radiation based method for locating and differentiating sentinel nodes. The method described is used to identify a sentinel lymph node located within a grouping of regional nodes at a lymph drainage basin connected to neoplastic tissue. A radionucleide is injected near the neoplastic tissue and migrates along a lymph duct toward the drainage basin containing the sentinel node. A hand-held, radiation detection probe is moved along the lymph duct while the operator observes a graphical readout of count rate amplitudes to determine when the probe is aligned with the duct. The region containing the sentinel node is identified when the count rate at the probe substantially increases. Following incision, the probe is maneuvered using a sound output to establish increasing count rate thresholds. The probe is then moved incrementally until the probe is adjacent to the sentinel node, which then may be surgically removed. The visual and audio signals used by the surgeon are generated by the signal processing portion of the radiation detection system, which may be referred to as the control unit control unit. The control unit is connected to the handheld probe to form the radiation detector system.
The success of using a method such as disclosed in U.S. Pat. No. 5,732,704 depends upon the reliability of the hand-held radiation detection probe and the calibration between the probe and the control unit. The probe generally operates at room temperature and is designed to detect very low levels of gamma radiation. The gamma radiation emitted from the sentinel node may be masked by background noise such as cosmic radiation, thermal noise, and capacitively or piezoelectrically induced noise resulting from manipulation of the probe itself. One function of the control unit is to filter gamma radiation emitted by the radio tagged tissue from background noise and other sources of gamma radiation, including Compton scatter.
Gamma ray detection probes may include a high-Z semiconductor (such as CdZnTe or CdTe) or a scintillation crystal such as sodium-iodide (NaI) which is coupled with a small photo multiplier tube. U.S. patent application Ser. No. 09/066,545, filed on Apr. 24, 1998, now abandoned, describes a relatively low-cost radiation detection probe. The probe integrates a silicon photodiode detector (with or without a scintillation assembly) with amplifiers, interface electronics, and radiation shielding, into one compact radiation probe assembly. The probe assembly uses relatively low voltages, has relatively few electrical connections, is relatively easy to manufacture, and is low-cost. The disclosed radiation detection probe is particularly useful for detecting radionucleides during lymphatic mapping and localization of a sentinel node.
Despite recent advances to lower the cost of manufacture and use of radiation detection probes, it is still necessary to insure that the electronic signal generated by the probes are correctly interpreted by the control unit. Careful attention to manufacturing tolerances and the use of specially selected electronic components may ensure adequate calibration between probes and adequate stability after the probes leave the factory, thus ensuring that the output for a given input is relatively constant across a selection of probes and relatively stable over time. Of course, such manufacturing tolerances and special electronics add significant cost. Lower cost probes, on the other hand, may be manufactured to wider tolerances and utilize less expensive electronics, making them less consistent probe to probe and more likely to lose calibration after leaving the factory. One particularly important characteristic of radiation detection probes is the signal output level generated by a predetermined signal input level. For example, one probe may generate an electronic pulse output of 5.1 volts when a gamma ray having an energy level of 140.5 kilo electron Volts (keV) is detected. An energy level of 140.5 keV is typical of a gamma ray photon generated by Technetium-99. However, because of a number of factors, including manufacturing tolerances and variations in electronic component characteristics, an identically manufactured probe may generate an electronic pulse of 4.9 volts when detecting a gamma ray photon having an energy level of 140.5 keV. Further, even if the output of a particular probe is within acceptable tolerances at the factory, the output signal level may shift over time. When using probes which vary over time or from probe to probe, the control units must, therefore, be calibrated using a known radioactive source so that the electronic output signals from the probe are correctly interpreted by the radiation detection system.
Radiation detection systems are typically calibrated against a radioisotope which has a known peak energy level. This may be accomplished by, for example, calibrating each radiation detection system periodically in a biomedical lab. The probe is held near a radioisotope having a known, characteristic, gamma radiation energy level. Each gamma ray photon emitted by the radioisotope represents a singular radioactive event and each gamma ray photon has an energy level measurable in kilo electron volts (keV). Each such gamma ray photon or radioactive event which is detected by a probe may be referred to as a count. Upon detecting gamma ray photons, the probe generates a series of electric pulses, each pulse having a voltage which is proportional to the energy level of a gamma ray photon. Since the probe is positioned directly adjacent a radioactive source emitting gamma ray photons of a known energy level, the number of counts associated with that energy level would be far higher than the number of counts from other sources such as background radiation or Compton scatter. Thus, with the probe positioned near a known source, the control unit may be adjusted to calibrate the system by identifying the probe output signal (e.g. voltage) having the highest number of occurrences within a predetermined time period or by accepting a predetermined number of counts and identifying the output signal (e.g. voltage) associated with the largest number of counts. The output signal associated with the largest number of counts may then be interpreted to represent the energy level of the calibration radiation source. In order to calibrate a probe properly a statistically significant number of counts must be used. The predominant pulse height, also called the peak pulse height, can be derived from the recorded spectrum of pulses. The peak pulse height is interpreted by the radiation detection system to correspond to the known, characteristic energy level of the radioisotope used for calibration. Once the peak pulse height has been identified for a particular probe, the control unit input window may be set to allow that signal to pass while filtering out other signals such as noise or Compton scatter.
Normally, probes are designed and manufactured to have a predetermined output signal level for a count of a predetermined energy level. Unfortunately, a probe can lose calibration between the time it is calibrated in a lab and the time it is used on a patient. Calibration loss (drift) can also occur due to mishandling of delicate probes or during prolonged storage periods. In addition, the radioisotopes typically used in the calibration lab are not always the same as those used in the surgical patient (it is desired to inject a radioisotope with a short half-life into the patient, whereas the half-life of the radioisotope used in the calibration lab is preferably long so that it can be used over an extended period of time). Thus, the energy level of the radioisotope used to calibrate the radiation detection system may be different then the peak energy level of the radionucleide injected into the patient. Therefore, in current systems it may be necessary to provide some means for extrapolating the results of the laboratory calibration to the actual surgical situation. Although the individual contributions probe drift, control unit drift, probe damage, and using calibration radioisotopes are typically small, it is desirable to reduce or eliminate them altogether. It would, therefore, be desirable to calibrate the control unit to the output of a particular probe immediately before its use, and preferably with the same radionucleide used in the patient. The latter approach would be practical if the physician operator could perform the calibration immediately prior to the procedure. The physician operator, however, often does not have the expertise of a nuclear imaging technician, nor is the physician operator working in the controlled conditions of a biomedical laboratory. What is needed is calibration method which could be easily performed by the physician operator in the operating room. The calibration method could be made suitable for use by physician operators by automating many of the steps and by providing appropriate feedback signals to the operator in order to properly position the probe during the calibration procedure. It would, therefore, be advantageous if the radiation detection system could be calibrated using the radionucleide injection site. In particular, the injection site in a sentinel node procedure may be suitable for use as a calibration source because most of the injected radionucleide remains at the injection site for many hours after injection; the lymphatic system drains a relatively small amount during that time. Alternatively, it would be advantageous to design a calibration method which used a separate radioactive source available to the physician operator, such as the radionucleide in the administration syringe which is available to the physician operator immediately prior to the surgery.
In addition to the need for the radiation detection system to be properly calibrated, a filter is still required to discern radioactive emissions of the radionucleide in suspect tissue from background radiation. This background radiation, results predominately from Compton scattering. Compton scattering (or scatter) results from the interaction of gamma ray photons with electrons of body tissues. The scattered gamma ray photons have energies ranging from slightly below the full energy gamma ray photons down to and below typical x-ray energies (the xe2x80x9cCompton continuumxe2x80x9d). The apparent points of origin of these Compton scattered gamma ray photons have only a limited relationship to the site from which the unscattered, full energy gamma ray photons originated, and therefore have little relationship to the location of the tissue of interest. When using the method disclosed in U.S. Pat. No. 5,732,704, much of this Compton scatter comes from the radionucleide concentrated around the injection site in the patient, and this radiation can obscure the gamma radiation emitted by the radionucleide that has migrated to the sentinel node. Discerning the gamma radiation emitted by the radionucleide in the sentinel node from all other radiation sources reliably and consistently for each surgical patient is a highly desired objective of the surgeon. Therefore it is desirable to be able to discriminate between those gamma ray photons having energies close to that of the full-energy gamma ray photon and background radiation. Therefore, it is desirable to utilize a filter or window within the control unit which eliminates the probe output signals representative of background radiation. Normally the filter output includes the probe output signal levels representative of a full energy gamma ray photon and excludes probe output signal levels which are representative of background radiation, including undesirable radiation resulting from Compton scattering, such filters are known in the art.
The prior art discloses radiation detection devices that remove background radiation using xe2x80x9cwindowing techniquesxe2x80x9d in order to discern and process the gamma radiation emitted by the radionucleide concentrated in suspect tissue. U.S. Pat. No. 5,694,933 issued to Madden et al on Dec. 9, 1997 discloses an apparatus having a hand-held probe, a signal processor, and a multichannel control unit (MCA) to identify a peak energy level, to set manually a window of energy levels, and to perform a variety of other functions.
Another phenomenon associated with radiation detection systems is commonly known in the art as xe2x80x9cpileupxe2x80x9d. Pileup occurs when the frequency of counts impinging on the forward window of the probe is higher than the response rate of the radiation detection system, especially the crystal and detector portion of the probe. Thus the system is unable to detect and process each count individually. As a result, a multiplicity of counts emitted by an especially xe2x80x9chotxe2x80x9d radiation source may be detected as a smaller number of counts having a higher energy level. Pileup phenomena are of two general types, which have somewhat different effects on pulse height measurements. The first type is known as tail pileup and involves the superposition of pulses on the long-duration tail from a preceding pulse. A second type of pileup is called peak pileup and occurs when two pulses are sufficiently close together so that they are treated as a single pulse by the radiation detection system. These types of pileup lead to distortions of the recorded pulse height spectrum and can cause a misinterpretation of the emissions of the radionucleide during both the calibrating of the probe and the locating of the sentinel node. A detailed description of pileup is provided in Radiation Detection and Measurement, by Glenn F. Knoll, pages 610-612, publisher John Wiley and Sons, Inc, (hereinafter Knoll). Knoll further describes electronic and statistical means for xe2x80x9cpileup rejectionxe2x80x9d (pages 612-620) in order to reduce but not totally eliminate the problems associated with pileup.
One method for reducing the effects of pileup is to position the probe farther away from the radioactive source. Radiation intensity is inversely proportional to the square of the distance from the radiation source. Therefore, the quantity of gamma ray photons from a particular source which impinge on the receiving window of the probe is reduced by moving the probe away from the source. It is difficult to properly calibrate the probe if it is not positioned correctly with respect to the calibration source. It would, therefore, be advantageous during calibration if a high-count, feedback signal is provided to the operator when the count frequency is high enough to result in a significant pileup distortion of the recorded spectrum. Then the operator may quickly reposition the probe-receiving window farther away from the radioactive source.
Another situation that may occur during calibration of a radiation detection system is when the operator does not position the receiving window of the probe close enough to the radioactive source. In accord with the inverse square law for radiation propagation, the resulting count frequency may be very low. If the count frequency is very low, the time required to detect a statistically significant number of counts may be high (several seconds). When count frequency is so low that it would take a significant length of time to collect the required number of counts, it would be advantageous to provide a low-count, feedback signal to the operator. Then the probe could be repositioned closer to the radioactive source. Furthermore, if the low-count feedback signal is generated when the count frequency is less than desired, and a high-count feedback signal is generated when count frequency is more than desired, then the operator is aided in positioning the probe receiving window the correct distance from the radioactive source. Finally, if a third feedback signal is generated when a desired count frequency is obtained (neither too high or too low), it would be even easier for the operator to correctly position the probe relative to the radiation source. Thus, it would be advantageous to design a radiation detection system having a calibration mode wherein the physician operator is assisted in positioning the probe during calibration of the radiation detection system. In such a system, the physician operator could calibrate the system during or immediately prior to initiating a surgical procedure.
The calibration positioning method described herein may be combined with an automatic windowing method for determining an energy acceptance window in order to reduce the effects of background radiation. An operator could use the combined methods to calibrate the radiation detection system using the injection site in the patient of the radionucleide as the radiation source. By using such a calibration method on the injection site, it would be practical to use low cost probes that generate a wide range of pulse magnitudes for counts of a given energy level. Also, the effects of calibration error due to electronic drift within the radiation detection system, damage to the probe during handling, and calibration on a different radioisotope may be diminished or avoided.
The present invention is a method of calibrating a radiation detection system by adjusting the parameters of a control unit for each probe attached to the control unit. The method described provides a reliable and practical way of using low cost radiation detection probes having variations in probe gain where probe gain is the ratio of voltage generated at the probe output for a given energy input. That is, probe gain is the factor which relates the energy deposited in a probe by a gamma ray photon to the voltage generated at the output of the probe where the voltage is generated as a result of the photon striking the detector input.
A method according to the present invention uses feedback to ensure that the user has an appropriate count rate to accurately calibrate the system. The method comprises the following steps: An operator positions a radiation detection probe near a calibration radiation source, such as the surgical injection site of a radionucleide. Then the radiation detection probe generates a plurality of electronic pulses, each of the electronic pulses having a pulse magnitude proportional to the energy level of a count. Next the radiation detection system calculates a pulse frequency. The radiation detection system generates a low-count feedback signal when the pulse frequency is less than a predetermined low-count frequency. The radiation detection system generates a high-count feedback signal when the pulse frequency is greater than a predetermined high-count frequency. The radiation detection system generates an optimal-count feedback signal when the pulse frequency is greater than or equal to the predetermined low-count frequency and is less than or equal to the predetermined high-count frequency. Each of the feedback signals may be an audible feedback signal, a visual feedback signal, or both. When the probe is positioned such that the optimal feedback signal is generated, the control unit begins to collect and record outputs from the probe. The control unit continues to collect and record output signals from the probe until a statistically significant number of output pulses have been recorded. The control unit then categorizes the plurality of pulses into a plurality of pulse magnitude ranges. Next the control unit determines the number of pulses categorized in each of the plurality of pulse magnitude ranges. The control unit then identifies the pulse magnitude range containing the most pulses and labels that range the peak pulse magnitude range. Then the system assigns an energy level to each of the plurality of pulse magnitude ranges. The energy level assigned to the peak pulse magnitude may be referred to as the characteristic energy level and is generally the characteristic energy level of gamma ray photons emitted by the calibration source. Once calibrated, the control unit may use the characteristic energy level for a plurality of purposes. For example, the control unit may determine the lower cut off limits of an energy acceptance filter which includes the characteristic energy level but excludes pulses representative of gamma ray photons having energy levels below a predetermined threshold energy level. The threshold energy level generally has a predetermined relationship to the characteristic energy level. The radiation detection system thereafter processes only pulses corresponding to an energy level passed by the energy acceptance filter. In another embodiment of the present invention, the energy acceptance filter also passes pulse representative of gamma ray photons having energy levels up to a highest energy level having a second predetermined relationship to the characteristic energy level.